diff --git a/CMakeLists.txt b/CMakeLists.txt
index a2e85aa..3055ee5 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -15,6 +15,9 @@ option(BUILD_DX11 "Build with DirectX11 Runtime support" ON)
 # Includes the functions for creating Direct3D 12 resources at runtime
 option(BUILD_DX12 "Build with DirectX12 Runtime support" ON)
 
+# Enable the use of OpenMP for software BC6H/BC7 compression
+option(BC_USE_OPENMP "Build with OpenMP support" ON)
+
 set(CMAKE_CXX_STANDARD 14)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
 set(CMAKE_CXX_EXTENSIONS OFF)
@@ -107,6 +110,8 @@ if ( CMAKE_CXX_COMPILER_ID MATCHES "Clang" )
     set(WarningsLib "-Wpedantic" "-Wextra")
     target_compile_options(${PROJECT_NAME} PRIVATE ${WarningsLib})
 
+    # OpenMP is not supported for clang for Windows by default
+
     set(WarningsEXE ${WarningsLib} "-Wno-c++98-compat" "-Wno-c++98-compat-pedantic" "-Wno-switch" "-Wno-switch-enum" "-Wno-language-extension-token" "-Wno-missing-prototypes")
     target_compile_options(texassemble PRIVATE ${WarningsEXE})
     target_compile_options(texconv PRIVATE ${WarningsEXE})
@@ -118,6 +123,11 @@ if ( CMAKE_CXX_COMPILER_ID MATCHES "MSVC" )
     target_compile_options(texconv PRIVATE /permissive- /Zc:__cplusplus)
     target_compile_options(texdiag PRIVATE /permissive- /Zc:__cplusplus)
 
+    if(BC_USE_OPENMP MATCHES ON)
+        target_compile_options(${PROJECT_NAME} PRIVATE /openmp /Zc:twoPhase-)
+        target_compile_options(texconv PRIVATE /openmp /Zc:twoPhase-)
+    endif()
+
     set(WarningsEXE "/wd4061" "/wd4062" "/wd4365" "/wd4668" "/wd4710" "/wd4820" "/wd5039" "/wd5045")
     target_compile_options(texassemble PRIVATE ${WarningsEXE})
     target_compile_options(texconv PRIVATE ${WarningsEXE})
diff --git a/DirectXTex/DirectXTexCompress.cpp b/DirectXTex/DirectXTexCompress.cpp
index c987183..a94ae0b 100644
--- a/DirectXTex/DirectXTexCompress.cpp
+++ b/DirectXTex/DirectXTexCompress.cpp
@@ -250,12 +250,12 @@ namespace
             assert((y >= 0) && (y < int(image.height)));
 
             size_t rowPitch = image.rowPitch;
-            const uint8_t *pSrc = image.pixels + (y*rowPitch) + (x*sbpp);
+            const uint8_t *pSrc = image.pixels + (size_t(y)*rowPitch) + (size_t(x)*sbpp);
 
-            uint8_t *pDest = result.pixels + (nb*blocksize);
+            uint8_t *pDest = result.pixels + (size_t(nb)*blocksize);
 
-            size_t ph = std::min<size_t>(4, image.height - y);
-            size_t pw = std::min<size_t>(4, image.width - x);
+            size_t ph = std::min<size_t>(4, image.height - size_t(y));
+            size_t pw = std::min<size_t>(4, image.width - size_t(x));
             assert(pw > 0 && ph > 0);
 
             ptrdiff_t bytesLeft = pEnd - pSrc;
@@ -268,19 +268,19 @@ namespace
 
             if (ph > 1)
             {
-                bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft - rowPitch));
+                bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft) - rowPitch);
                 if (!_LoadScanline(&temp[4], pw, pSrc + rowPitch, bytesToRead, format))
                     fail = true;
 
                 if (ph > 2)
                 {
-                    bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft - rowPitch * 2));
+                    bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft) - rowPitch * 2);
                     if (!_LoadScanline(&temp[8], pw, pSrc + rowPitch * 2, bytesToRead, format))
                         fail = true;
 
                     if (ph > 3)
                     {
-                        bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft - rowPitch * 3));
+                        bytesToRead = std::min<size_t>(rowPitch, size_t(bytesLeft) - rowPitch * 3);
                         if (!_LoadScanline(&temp[12], pw, pSrc + rowPitch * 3, bytesToRead, format))
                             fail = true;
                     }